Devices for reconstructing video holograms using acousto-optical modulators (AOM) are known from prior art, as detailed in, for instance, U.S. Pat. No. 5,172,251 issued to Benton et al. on Dec. 15, 1992 entitled “Three dimensional display system,” the contents of which are hereby incorporated by reference. Such acousto-optical modulators transform electric signals into optical wave fronts, which are recomposed in a video frame using deflection mirrors to form two-dimensional holographic areas. A scene visible for the viewer is reconstructed from the individual wave fronts using further optical elements. The optical means used, such as lenses and deflection elements, have the dimensions of the reconstructed scenes. Due to their great depth, these elements are voluminous and heavy. It is difficult to miniaturize them, so that their range of applications is limited.
Another way to generate large video holograms is the so-called “tiling method”, using computer-generated holograms (CGH). This method is described in for instance PCT patent publication WO 00/75698 of PCT Patent Application PCT/GB2000/001901 filed on May 18, 2000 by Pain et. al entitled “Holographic Displays” and in U.S. Pat. No. 6,437,919 B1 issued to Brown, et al. on Aug. 20, 2002 “System for the production of a dynamic image for display”, both of which are hereby incorporated by reference. In the tiling method, small CGHs having a small pitch are created using an optical system. In the first step of the method, the required information is written to fast matrices that have a small pitch, such as electronically addressable spatial light modulators (EASLM). These fast matrices are then reproduced on to a portion of a suitable holographic medium. A large video hologram is composed of the tiled replicas of the fast matrices. Usually, an optically addressable spatial light modulator (OASLM) is used as holographic medium. In a second step, the composed video hologram is reconstructed with coherent light in transmission or reflection.
In the CGH with controllable openings arranged in a matrix or in an otherwise regular pattern as described in, for instance, PCT publications number WO 01/95016 A1 of PCT patent application PCT/GB2001/002302 filed on May 24, 2001 by Payne et al. entitled “Computation Time Reduction for Three-Dimensional Displays” or in Fukaya et al., “Eye-position tracking type electro-holographic display using liquid crystal devices”, Proceedings of EOS Topical Meeting on Diffractive Optics, 1997, the diffraction on small openings is taken advantage of for encoding the scenes. The wave fronts emerging from the openings converge in object points of the three-dimensional scene before they reach the viewer. The smaller the pitch, and thus the smaller the openings in the CGHs, the greater is the diffraction angle, i.e. the viewing angle. Consequently, with these known methods enlarging the viewing angle means to improve the resolution.
As is generally known, in Fourier holograms the scene is reconstructed as a direct or inverse Fourier transform of the hologram in a plane. This reconstruction is continued periodically at a periodicity interval, the extension of the periodicity interval being inversely proportional to the pitch in the hologram.
If the dimension of the reconstruction of the Fourier hologram exceeds the periodicity interval, adjacent diffraction orders will overlap. As the resolution is gradually decreased, i.e. as the pitch of the openings increases, the edges of the reconstruction will be increasingly distorted by overlapping higher diffraction orders. The usable extent of the reconstruction is thus gradually limited.
If greater periodicity intervals and thus greater viewing angles are to be achieved, the required pitch in the hologram comes closer to the wavelength of the light. Then, the CGHs must be sufficiently large in order to be able to reconstruct large scenes. These two conditions require a large CGH having a great number of openings. However, this is currently not feasible in the form of displays with controllable openings, as discussed in, for instance, U.S. Pat. No. 6,831,678 issued to Travis on Dec. 14, 2004 entitled “Autostereoscopic display”, the contents of which are hereby incorporated by reference. CGH with controllable openings only measure one to several inches, with the pitches still being substantially greater than 1 μm.
The two parameters, pitch and hologram size, are characterized by the so-called space-bandwidth product (SBP) as the number of openings in the hologram. If the reconstruction of a CGH with controllable openings that has a width of 50 cm is to be generated so that a viewer can see the scene at a distance of 1 m and in a 50-cm-wide horizontal viewing window, the SBP in horizontal direction is about 0.5·106. This corresponds to 500,000 openings at a distance of 1 μm in the CGH. Assuming an aspect ratio of 4:3, 375,000 openings are required in the vertical direction. Consequently, the CGH comprises 3.75·1011 openings, if three color sub-pixels are taken into consideration. This number will be tripled if the fact is taken into account that the CGH with controllable openings usually only allows the amplitudes to be affected. The phases are encoded taking advantage of the so-called detour phase effect, which requires at least three equidistant openings per sampling point. SLM having such a great number of controllable openings are hitherto unknown.
The hologram values must be calculated from the scenes to be reconstructed. Assuming a color depth of 1 Byte for each of the three primary colors and a frame rate of 50 Hz, a CGH requires an information flow rate of 50·1012=0.5·1014 Byte/s. Fourier transformations of data flows of this magnitude exceed the capabilities of today's computers by far and thus do not allow holograms to be calculated based on local computers. However, transmitting such an amount of data through data networks is presently unfeasible for normal users.
In order to reduce the enormous number of computations it has been proposed not to calculate the entire hologram, but only such parts of it that can be seen directly by the viewer, or such parts that change. The kind of hologram which consists of addressable sub-regions, such as the above-mentioned “tiling hologram”, is disclosed in the above-mentioned patent specification WO 01/95016 A1. Starting point of the calculations is a so-called effective exit pupil, the position of which can coincide with the eye pupil of the viewer. The image is tracked as the viewer position changes by continuous recalculation of the hologram part that generates the image for the new viewer position. However, this partly nullifies the reduction in the number of computations.
The disadvantages of the known methods can be summarized as follows: Arrangements with acousto-optical modulators are too voluminous and cannot be reduced to dimensions known from state-of-the-art flat displays; video holograms generated using the tiling method are two-stage processes which require enormous technical efforts and which cannot easily be reduced to desktop dimensions; and arrangements based on SLM with controllable openings are too small to be able to reconstruct large scenes. There are currently no large controllable SLM with extremely small pitches, which would be needed for this, and this technology is further limited by the computer performance and data network bandwidth available today.